mwang@watmath.UUCP (mwang) (05/27/85)
DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
COMPUTER SCIENCE COLLOQUIUM
- Thursday, June 6, 1985.
Dr. I. Bruha of Acadia University will speak on ``One
System of Structure Learning Implemented in PROLOG.''
TIME: 3:30 PM
ROOM: MC 6091A (Please Note)
ABSTRACT
The author will firstly discuss the main approaches to
representation of patterns and different types of
learning systems: feature, syntax, structural and
rule-based approaches. Three concrete examples of
structure learning systems (Winston, Shapiro, Bratko),
all being implemented in PROLOG, will be explained.
Afterwards, the author's system of a knowledge acquisi-
tion for an expert system will be presented. There
exist many approaches to knowledge acquisition; one
possibility is to utilize PROLOG and its deduction pro-
perty for the structure learning. PROLOG can be used
both for acquisition of production rules from examples
and for testing.
Expert systems usually involve fuzzy information but
the language PROLOG does not process numbers in a good
manner. Therefore the author has implemented an
extended version PROLOG, called PROLOGTRAN. The learn-
ing system, implemented in this language, can easily
process both structural and numerical information.mwang@watmath.UUCP (mwang) (05/28/85)
DEPARTMENT OF COMPUTER SCIENCE
UNIVERSITY OF WATERLOO
COMPUTER SCIENCE COLLOQUIUM
- Wednesay, June 5, 1985.
Dr. I. Bruha of Acadia University will speak on ``One
System of Structure Learning Implemented in PROLOG.''
TIME: 3:30 PM
ROOM: MC 5158 (Please Note)
ABSTRACT
The author will firstly discuss the main approaches to
representation of patterns and different types of
learning systems: feature, syntax, structural and
rule-based approaches. Three concrete examples of
structure learning systems (Winston, Shapiro, Bratko),
all being implemented in PROLOG, will be explained.
Afterwards, the author's system of a knowledge acquisi-
tion for an expert system will be presented. There
exist many approaches to knowledge acquisition; one
possibility is to utilize PROLOG and its deduction pro-
perty for the structure learning. PROLOG can be used
both for acquisition of production rules from examples
and for testing.
Expert systems usually involve fuzzy information but
the language PROLOG does not process numbers in a good
manner. Therefore the author has implemented an
extended version PROLOG, called PROLOGTRAN. The learn-
ing system, implemented in this language, can easily
process both structural and numerical information.
Coffee and refreshments will be served at 3 PM.